Bacterial DNA has immune stimulatory effects to activate B cells and natural killer cells, but vertebrate DNA does not (Tokunaga, T., et al., 1988. Jpn. J. Cancer Res. 79:682-686; Tokunaga, T., et al., 1984, JNCI 72:955-962; Messina, J. P., et al., 1991, J. Immunol. 147:1759-1764; and reviewed in Krieg, 1998, In: Applied Oligonucleotide Technology, C. A. Stein and A. M. Krieg, (Eds.), John Wiley and Sons, Inc., New York, N.Y., pp. 431-448). It is now understood that these immune stimulatory effects of bacterial DNA are a result of the presence of unmethylated CpG dinucleotides in particular base contexts (CpG motifs), which are common in bacterial DNA, but methylated and underrepresented in vertebrate DNA (Krieg et al, 1995 Nature 374:546-549; Krieg, 1999 Biochim. Biophys. Acta 93321:1-10).
The immune stimulatory effects of bacterial DNA can be mimicked with synthetic oligodeoxynucleotides (ODN) containing these CpG motifs. Such CpG ODN have highly stimulatory effects on human and murine leukocytes, inducing B cell proliferation; cytokine and immunoglobulin secretion; natural killer (NK) cell lytic activity and IFN-γ secretion; and activation of dendritic cells (DCs) and other antigen presenting cells to express costimulatory molecules and secrete cytokines, especially the Th1-like cytokines that are important in promoting the development of Th1-like T cell responses. These immune stimulatory effects of native phosphodiester backbone CpG ODN are highly CpG specific in that the effects are essentially abolished if the CpG motif is methylated, changed to a GpC, or otherwise eliminated or altered (Krieg et al, 1995 Nature 374:546-549; Hartmann et al, 1999 Proc. Natl. Acad. Sci USA 96:9305-10). Phosphodiester CpG ODN can be formulated in lipids, alum, or other types of vehicles with depot properties or improved cell uptake in order to enhance the immune stimulatory effects (Yamamoto et al, 1994 Microbiol. Immunol. 38:831-836; Gramzinski et al, 1998 Mol. Med. 4:109-118).
In early studies, it was thought that the immune stimulatory CpG motif followed the formula purine-purine-CpG-pyrimidine-pyrimidine (Krieg et al, 1995 Nature 374:546-549; Pisetsky, 1996 J. Immunol. 156:421-423; Hacker et al., 1998 EMBO J. 17:6230-6240; Lipford et al, 1998 Trends in Microbiol. 6:496-500). However, it is now clear that mouse lymphocytes respond quite well to phosphodiester CpG motifs that do not follow this “formula” (Yi et al., 1998 J. Immunol. 160:5898-5906) and the same is true of human B cells and dendritic cells (Hartmann et al, 1999 Proc. Natl. Acad. Sci USA 96:9305-10; Liang, 1996 J. Clin. Invest. 98:1119-1129).
Several past investigators have looked at whether the nucleotide content of ODN may have effects independently of the sequence of the ODN. Interestingly, antisense ODN have been found to be generally enriched in the content of GG, CCC, CC, CAC, and CG sequences, while having reduced frequency of TT or TCC nucleotide sequences compared to what would be expected if base usage were random (Smetsers et al., 1996 Antisense Nucleic Acid Drug Develop. 6:63-67). This raised the possibility that the over-represented sequences may comprise preferred targeting elements for antisense oligonucleotides or visa versa. One reason to avoid the use of thymidine-rich ODN for antisense experiments is that degradation of the ODN by nucleases present in cells releases free thymidine which competes with 3H-thymidine which is frequently used in experiments to assess cell proliferation (Matson et al., 1992 Antisense Research and Development 2:325-330).